mirror of
https://github.com/lux-apps/papers.git
synced 2026-10-30 18:15:21 +00:00
zeekay
f4ee7a0525
Bylines, paper authorship, correspondence addresses and the founder blurb under each post named an individual. They name the organisation that published the work instead: the brand on posts and pages, the research byline on papers, and research@ in place of a personal address. Both are what these repositories already used elsewhere — nothing new was coined. Co-authors are untouched. Where a paper named someone else alongside, they stay named; only the one byline changed. Bibliography entries, running heads, PDF metadata and affiliations move with it, so a citation still resolves to whoever the cited paper names. Sample and fixture data uses the placeholder name these tests already use elsewhere, so no assertion changed meaning.
86 lines
3.2 KiB
TeX
86 lines
3.2 KiB
TeX
\documentclass[11pt]{article}
|
|
\usepackage{../shared/luxcover}
|
|
\usepackage{amsmath,amssymb}
|
|
\usepackage{graphicx}
|
|
\usepackage{booktabs}
|
|
\usepackage{hyperref}
|
|
\usepackage{listings}
|
|
\usepackage{xcolor}
|
|
\usepackage{tikz}
|
|
\usepackage{pgfplots}
|
|
\pgfplotsset{compat=1.18}
|
|
\usepackage{siunitx}
|
|
\usepackage{multirow}
|
|
\usepackage{subcaption}
|
|
\usepackage{algorithm}
|
|
\usepackage{algpseudocode}
|
|
\usetikzlibrary{arrows.meta,positioning,calc,shapes.geometric,fit}
|
|
|
|
\sisetup{group-separator={,}}
|
|
|
|
\definecolor{luxblue}{RGB}{30,64,175}
|
|
\definecolor{luxgold}{RGB}{234,179,8}
|
|
\definecolor{gpugreen}{RGB}{22,163,74}
|
|
\definecolor{cpured}{RGB}{220,38,38}
|
|
\definecolor{cppblue}{RGB}{37,99,235}
|
|
\definecolor{gogreen}{RGB}{0,125,156}
|
|
\definecolor{fhepurple}{RGB}{124,58,237}
|
|
\definecolor{pqorange}{RGB}{234,88,12}
|
|
\definecolor{nttcyan}{RGB}{6,182,212}
|
|
|
|
\title{cevm: A GPU-Native Ethereum Virtual Machine\\with Post-Quantum Cryptography and FHE Acceleration}
|
|
\author{
|
|
Lux Research\thanks{Corresponding author: [email protected]}\\
|
|
\textit{Lux Industries, Inc.}\\
|
|
\texttt{[email protected]}
|
|
}
|
|
\date{April 2025}
|
|
|
|
\begin{document}
|
|
|
|
\luxcoverpage
|
|
|
|
\begin{abstract}
|
|
We present \textbf{cevm}, a GPU-native Ethereum Virtual Machine that executes the entire
|
|
block processing pipeline---transaction validation, signature recovery, Block-STM parallel
|
|
execution, opcode dispatch, and state root computation---on GPU hardware. Measured on
|
|
Apple M1 Max (32-core Metal GPU, 64\,GB unified memory), cevm achieves 85.9\,M opcodes/sec
|
|
on the GPU EVM kernel (4.85$\times$ over CPU), 20.9\,Ggas/s on the C++ interpreter with
|
|
full state (5.1$\times$ over Go), and 7.1$\times$ speedup on parallel ecrecover across 10 CPU
|
|
cores.
|
|
|
|
Beyond EVM acceleration, cevm provides GPU-native implementations of all three NIST
|
|
post-quantum FIPS standards (ML-DSA, ML-KEM, SLH-DSA), Lux-specific Pulsar signatures,
|
|
threshold protocols (FROST, CGGMP21), and TFHE fully homomorphic encryption---sharing
|
|
the Number Theoretic Transform (NTT) as a unified GPU primitive across all lattice-based
|
|
operations.
|
|
|
|
The system comprises 30 Metal compute shaders (production, Apple Silicon), 30 CUDA kernels
|
|
(ported from Metal, ready for H100 CI), and 30 WGSL shaders (ported from Metal, validated
|
|
with naga)---90 shader files total with zero stubs remaining across three backends: Metal
|
|
(production), CUDA (ready), and Dawn/WebGPU (ready), plus CPU SIMD fallback on all
|
|
platforms. The system is validated by 1{,}183 tests across C++, Rust, and Go. GPU memory
|
|
scales linearly at 73\,KB per transaction, supporting blocks of 800{,}000 transactions on
|
|
64\,GB unified memory.
|
|
\end{abstract}
|
|
|
|
\tableofcontents
|
|
\newpage
|
|
|
|
%% ============================================================================
|
|
|
|
\input{sections/01-introduction}
|
|
\input{sections/02-architecture-the-cevm-pipeline}
|
|
\input{sections/03-gpu-state-hash-table-design}
|
|
\input{sections/04-block-stm-on-gpu}
|
|
\input{sections/05-post-quantum-cryptography-acceleration}
|
|
\input{sections/06-fhe-on-gpu-encrypted-evm-computation}
|
|
\input{sections/07-benchmarks}
|
|
\input{sections/08-updated-benchmark-results}
|
|
\input{sections/09-comparison-with-external-gpu-evm-project}
|
|
\input{sections/10-cuda-backend-and-h100-projections}
|
|
\input{sections/11-test-coverage-and-validation}
|
|
\input{sections/99-conclusion}
|
|
|
|
\end{document}
|